The problem is they have both: lack of interest towards IPFS and $300 million weight to keep control over IPFS development. So doing all the necessary work to give up control and encourage quality independent implementations is out of the question.